Least Cost Strategies for Complying with New NOx Emissions Limits

Transcription

1 Least Cost Strateges for Complyng wth New NOx Emssons Lmts Internatonal Assocaton for Energy Economcs New England Chapter Presented by Assef A. Zoban Tabors Caramans & Assocates Cambrdge, MA January 25, 2000 January 25,

2 Presentaton Outlne Introducton Market Model-Mathematcal Formulaton Perfect Complance Over Complance Insghts from the Mathematcal Model Generators Bddng Behavor Investment Decson Crteron Dynamcs of Tradable Permts Prces Practcal Approach General Market Smulaton Methodology Emssons Modelng Applcaton to the Northeastern & Mdwest US Energy Markets Analyss of the Results Conclusons January 25,

3 Introducton In a Memorandum of Understandng (MOU), the 12-state Northeast Ozone Transport Regon (OTR) states volunteered to reduce emssons through nsttuton of a cap-and-trade program. Phase II of the MOU allocates allowances based on the less strngent of a 75% reducton and a reducton to 0.15lb/MMBtu (effectve 1999) Phase III scheduled to begn n 2003 In 1997, EPA ssued the State Implementaton Plan Call, whch requre 22 states n the Eastern US to submt plans to address the transport of ozone across state boundares. On May 1999, the D.C. Court of Appeals rulng suspended the mplementaton of the SIP Call. On December 17, 1999 ssued the Secton 126 fnal acton that requres 392 facltes n 12 states and DC to reduce annual emssons by a total of nearly 510,000 tons from 2007 levels. Affected sources are expected to comply by May 1, 2003 Each affected party wll partcpate n a federal NOx emssons capand-trade program January 25,

4 Objectves The recent court rulngs on EPA s NOx SIP Call ndcate that there s a strong need to quantfy the costs and benefts of NOx regulatons n the US. There has been speculaton that deregulatng the electrcty markets wll cause major harm to the Northeast regon by emssons from Mdwestern generaton. Identfy and evaluate complance optons Abatement technologes Tradable permts The effectveness of a tradable-permts markets n achevng effcent outcomes for envronmental emssons has not yet been fully modeled and analyzed for deregulated electrc power markets. January 25,

5 Complance Optons STRATEGY Invest Trade Retre? Fxed Cost Annual carryng cost of NOx abatement technology None Varable Cost (n energy bd) Operatng cost of abatement technology + Opportunty cost of used allowances Cost of purchased allowances + Opportunty cost of used allowances + Opportunty cost of lower dspatch January 25,

6 Market Model-Mathematcal Formulaton The combned energy and tradable permts markets can be smulated as a sngle mult-perod least-cost optmzaton problem wth demand balance and emssons budget constrants. Mn g, E r TotalCost T = t = 1 [ C ( g ( t)) + V ( E ) g ( t) ] + I ( E ) r r (1) Subject to: g ( t) Demand ( t) = :l(t) Energy Balance Constrant T t = 1 g ( t)( E A E r ) EmssonBudget :m Emssons Budget Constrant g ( t), E r 0 January 25,

7 Market Model-Mathematcal Formulaton Where: g (t) : Energy generated from unt at tme t. C ( g ( t)) : The generaton cost functon for unt at tme t,.e., cost of fuel and unt s varable operaton and mantenance cost. E A : Actual emsson rate for generaton unt before any abatement technology addton. We assume the emsson rate s fxed and ndependent of generaton. E r : Emsson rate reducton acheved by addng an abatement technology. V ( E r ) : Varable cost assocated wth reducng emssons from unt, by E r, we assume ths cost to be a lnear functon of E r, V ( E r ) = K E r. I ( E r ) : Fxed operatng and captal cost functon assocated wth emssons reductons, E r, over a perod T. We assume ths cost to be contnuos, convex and monotoncally ncreasng. λ (t) : Shadow prce of the energy balance constrant, or energy market-clearng prce at tme t. µ : Shadow prce of the emssons budget constrant, or market-clearng prce of tradable allowances. t [ 1, T ] : T s the set of ozone seasons, from May 1 st to September 30 th, over the average lfe expectancy of control technologes. [ 1, N ] : The set of all generators ncludng optmal (chosen) entry and retrement profle. January 25,

8 Market Model-Mathematcal Formulaton The Khun-Tucker condtons for the above optmzaton problem are: ' λ( t ) = C ( g ( t)) + µ ( E E ) + K ' I ( E r T t= 1 A ) + K g (t) = µ g(t) T t= 1 r E r, t (2) (3) µ 0, Wth complementary constrant: µ ( T t= 1 g ( t)( E A E r ) EmssonBudget) = 0 (4) January 25,

9 CASE I : Perfect Complance Total Emssons at Budget The market-clearng prce for the tradable allowances s the shadow prce of the emsson budget constrant, or the system cost reducton acheved by relaxng the emsson constrant by one per unt. The ncrease n market-clearng prce value s the cost of used tradable allowances and varable O&M costs assocated wth abatement technology. From equaton (3), for each unt, the total cost of tradng s equal to the ncremental cost of reducng emssons va abatement technologes, generators are ndfferent between nvestng or tradng (assumpton of contnuos nvestment functon). The tradable permt prce does not vary wth tme, whch rests on the assumpton that nvestments are made smultaneously, at whch tme the market acheves equlbrum. January 25,

10 CASE II : Over Complance Total Emssons wthn Budget The shadow prce of the budget constrant s zero, thus as shown n equaton (2a), the energy market-clearng prce s functon of margnal cost of the energy and control technology varable cost. Equaton (3a) shows that ths s not a feasble soluton snce the margnal cost of nvestment and the varable cost are both postve. Thus over-nvestment s not an optmal soluton for contnuous nvestment functon. However, n realty the market mght reach that level because of dscreteness and economes of scale n emsson control technologes. ' λ ( t ) = C ( g ( t)) + KE (2a) I ' ( E r T ) + K g (t) t= 1 January 25, r = 0 (3a)

11 Insghts from the Mathematcal Model Generators should bd ther margnal producton cost(.e., fuel cost plus tradng opportunty cost, plus any VOM assocated wth emsson reducton technologes). The energy market-clearng prce wll be set by the margnal unt(s) margnal producton cost. Generators should nvest n emsson reducton technologes as long as ther total cost of nvestment (captal and operatng) s less than the tradable permts cost. The tradable permts market-clearng prce wll exceed, equal, or be below the ncremental cost of emsson reducton n the case of under, perfect or over complance, respectvely. The ncremental cost of emsson reducton s related to the ncremental nvestment cost n reducton technology dvded by the total energy generated plus the technology VOM. January 25,

12 General Market Smulaton Methodology We utlzed GE-MAPS to model the electrc power generaton markets, n an teratve approach to solve the real verson of the above formulated problem. GE-MAPS s a securty-constraned least-cost chronologcal producton cost model. It s used to determne the locatonal energy market-clearng prces, the revenues, costs and proftablty of generaton unts. We used the most up to date data on load forecast, fuel prce, thermal unts avalablty (nuclear), thermal unts heat rates and fxed and operatng costs, transmsson constrants, and market rules. Why an teratve approach? Model capabltes to solve jont optmzaton of energy dspatch and nvestment decsons are not readly avalable. The generaton nvestment problem s solved separately n an teratve approach (new entry and retrements). January 25,

13 Emssons Modelng Assumptons Assume a perfect compettve market for tradable permts wth no transacton cost. Assume a cap-and-trade emsson reducton program wth budget constrants only (no unt or tme specfc constrants, and no bankablty). The cap-and-trade program s appled on a regonal (22-state, ncludng Northeast and Mdwest) bass rather than on a state by state bass. January 25,

14 Investment n Emsson Reducton - Algorthm STRATEGY Invest Trade Retre? Fxed Cost Annual carryng cost of NOx abatement technology None Varable Cost (n energy bd) Operatng cost of abatement technology + Opportunty cost of used allowances Cost of purchased allowances + Opportunty cost of used allowances + Opportunty cost of lower dspatch January 25,

15 Investment n Emsson Reducton - Algorthm 1. Start wth least-cost dspatch gnorng envronmental costs, determne unts generaton, revenues and costs. 2. Select a projected equlbrum tradng allowance prce, and compare the cost of tradng to the cost of nvestng (evaluate dfferent technologes), gven the performance level assumed n 1. Choose the opton that results n lower costs for each evaluated unt. 3. For those unts that opted to nvest, add the varable O&M of the selected technology to ther generaton bd. For all unts add the emsson opportunty costs as the tradable allowance prce tmes ther emsson rate (ether orgnal or post-nvestment). 4. Solve for least-cost dspatch wth the new unt margnal costs, determne unts generaton, revenues and costs, and total NO x emssons. 5. Check to see f total emssons are wthn budget. If yes, stop teratons, f no, go back to 2 (ncreasng the projected equlbrum allowance prce). January 25,

16 Applcaton to the Northeastern and Mdwest US Electrcty Markets Under the SIP Call, states were allocated budgets based on a NOx emsson rate of 0.15 lb/mmbtu and projected generaton levels. The total budget for the 22 states s 544,000 tons We consder the SIP Call n lght of the proposed and expected new entry nto the generaton markets n the Northeast and Mdwest. Assumed compettve market forces mprove avalablty of nuclear and fossl unts, and reduce operaton costs. January 25,

17 Impact on Northeast Markets Market Prces: n 2003, prces ncrease by up to 5% n PJM, 2-4% n NYPP and NEPOOL(relatve to absence of tghter lmts). However, the combned mpact of envronmental regulatons and new entry s to reduce the prces relatve to today. Investment cost: a very small ncremental cost assocated wth the Nox SIP Call was estmated (around $40 Mllon/year), because several nvestments have been made as part of Phase I of MOU n the OTR. Capacty Profle: sgnfcant new entry helps n dsplacng drter unts, and causes some retrements. The new entry sgnfcantly exceed the load growth and s more economc than many exstng unts. January 25,

18 Impact on Coal-Fred Generaton Unts 2,500 Heat Input (Mllons MMBtu) 2,000 1,500 1, NEPOOL NYPP PJM January 25,

19 Impact on Mdwest Electrcty Markets Market Prces: n 2003, prces ncrease by up to 15% n ECAR (relatve to absence of tghter lmts). However, the combned mpact of envronmental regulatons and new entry s to reduce the prces relatve to today. Investment cost: the cost assocated wth abatement technology assocated wth the SIP Call s sgnfcantly hgher than n the Northeast, and many more unts wll be mpacted (~$ 1 Bllon/year). The reason for ths hgher cost s the hgher porton of coal n the generaton mx n the Mdwest. Capacty Profle: sgnfcant new entry helps n dsplacng drter unts, and causes some retrements. The new entry sgnfcantly exceed the load growth and s more economc than many exstng unts. January 25,

20 Impact of Nox Emssons Tradng on ECAR Supply Curve Steam Coal - No NOx CCs - No NOx Steam Coal - Wth NOx CCs - Wth NOx 60 Prce ($/MWh) Wth NOx Tradng 20 No NOx Tradng Cumulatve Capacty (MW) January 25,

21 Conclusons The above proposed formulaton can be used by the ndustry to make nformed polcy decsons, and to evaluate the mpact of envronmental regulatons on market clearng prces of electrcty and the costs of emsson reducton for generators. The mpact of EPA s NOx SIP Call on energy market-clearng prces n the Northeastern and Mdwest US can be up to 5% n PJM and up to 15% n ECAR. The compettve entry wll reduce the ncremental cost assocated wth the NOx SIP Call. The analyss shows that the deregulaton of the electrc power markets and the envronmental regulatons can jon hands n reducng emssons from power plants. January 25,